Instructions to use sovasoft/zora-v1.11 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use sovasoft/zora-v1.11 with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("sovasoft/zora-v1.11", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- llama.cpp
How to use sovasoft/zora-v1.11 with llama.cpp:
Install (macOS, Linux)
curl -LsSf https://llama.app/install.sh | sh # Start a local OpenAI-compatible server with a web UI: llama serve -hf sovasoft/zora-v1.11:Q5_K_M # Run inference directly in the terminal: llama cli -hf sovasoft/zora-v1.11:Q5_K_M
Install from WinGet (Windows)
winget install llama.cpp # Start a local OpenAI-compatible server with a web UI: llama serve -hf sovasoft/zora-v1.11:Q5_K_M # Run inference directly in the terminal: llama cli -hf sovasoft/zora-v1.11:Q5_K_M
Use pre-built binary
# Download pre-built binary from: # https://github.com/ggerganov/llama.cpp/releases # Start a local OpenAI-compatible server with a web UI: ./llama-server -hf sovasoft/zora-v1.11:Q5_K_M # Run inference directly in the terminal: ./llama-cli -hf sovasoft/zora-v1.11:Q5_K_M
Build from source code
git clone https://github.com/ggerganov/llama.cpp.git cd llama.cpp cmake -B build cmake --build build -j --target llama-server llama-cli # Start a local OpenAI-compatible server with a web UI: ./build/bin/llama-server -hf sovasoft/zora-v1.11:Q5_K_M # Run inference directly in the terminal: ./build/bin/llama-cli -hf sovasoft/zora-v1.11:Q5_K_M
Use Docker
docker model run hf.co/sovasoft/zora-v1.11:Q5_K_M
- LM Studio
- Jan
- Ollama
How to use sovasoft/zora-v1.11 with Ollama:
ollama run hf.co/sovasoft/zora-v1.11:Q5_K_M
- Unsloth Desktop
- Pi
How to use sovasoft/zora-v1.11 with Pi:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf sovasoft/zora-v1.11:Q5_K_M
Configure the model in Pi
# Install Pi: npm install -g @earendil-works/pi-coding-agent # Add to ~/.pi/agent/models.json: { "providers": { "llama-cpp": { "baseUrl": "http://localhost:8080/v1", "api": "openai-completions", "apiKey": "none", "models": [ { "id": "sovasoft/zora-v1.11:Q5_K_M" } ] } } }Run Pi
# Start Pi in your project directory: pi
- Docker Model Runner
How to use sovasoft/zora-v1.11 with Docker Model Runner:
docker model run hf.co/sovasoft/zora-v1.11:Q5_K_M
- Lemonade
How to use sovasoft/zora-v1.11 with Lemonade:
Pull the model
# Download Lemonade from https://lemonade-server.ai/ lemonade pull sovasoft/zora-v1.11:Q5_K_M
Run and chat with the model
lemonade run user.zora-v1.11-Q5_K_M
List all available models
lemonade list
- Hermes Agent
How to use sovasoft/zora-v1.11 with Hermes Agent:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf sovasoft/zora-v1.11:Q5_K_M
Configure Hermes
# Install Hermes: curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash hermes setup # Point Hermes at the local server: hermes config set model.provider custom hermes config set model.base_url http://127.0.0.1:8080/v1 hermes config set model.default sovasoft/zora-v1.11:Q5_K_M
Run Hermes
hermes
- Atomic Chat
- OpenClaw
How to use sovasoft/zora-v1.11 with OpenClaw:
Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp # Start a local OpenAI-compatible server: llama serve -hf sovasoft/zora-v1.11:Q5_K_M
Configure OpenClaw
# Install OpenClaw: npm install -g openclaw@latest # Register the local server and set it as the default model: openclaw onboard --non-interactive --mode local \ --auth-choice custom-api-key \ --custom-base-url http://127.0.0.1:8080/v1 \ --custom-model-id "sovasoft/zora-v1.11:Q5_K_M" \ --custom-provider-id llama-cpp \ --custom-compatibility openai \ --custom-text-input \ --accept-risk \ --skip-health
Run OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"
Configure OpenClaw
# Install OpenClaw:
npm install -g openclaw@latest# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
--auth-choice custom-api-key \
--custom-base-url http://127.0.0.1:8080/v1 \
--custom-model-id "sovasoft/zora-v1.11:" \
--custom-provider-id llama-cpp \
--custom-compatibility openai \
--custom-text-input \
--accept-risk \
--skip-healthRun OpenClaw
openclaw agent --local --agent main --message "Hello from Hugging Face"๐ EN ยท SR ยท HR ยท BS ยท MK ยท SL ยท SQ ยท CNR ยท BG ยท EL ยท TR ยท RO ยท HU
Zora v1.11 โ an open, honest LLM for the Balkans & Southeast Europe
ะทะพัะฐ = โdawn". One to unite them all. โ by Sovasoft (ai.in.rs)
1 ยท What Zora is
Zora is an open 8B language model (built on Qwen3-8B) for 12 languages of the Balkans and Southeast Europe: Serbian, Croatian, Bosnian, Macedonian, Slovenian, Albanian, Montenegrin, Bulgarian, Greek, Turkish, Romanian, Hungarian.
Zora is not built to be the biggest model โ it is built to be honest, in-language, and multi-perspective:
- thinks in the target language instead of pivoting through English,
- shows several perspectives on contested topics instead of one national view,
- and above all: admits when it doesn't know instead of inventing facts.
2 ยท The development story (v1.0 โ v1.1 โ v1.11)
| Version | Languages | State |
|---|---|---|
| v1.0 | 6 | first public release |
| v1.1 | 12 | trained from scratch โ but hallucinated facts (invented book titles, wrong authors). Never released. |
| v1.11 | 12 | the honest fix: says โI don't know", searches when unsure, in-language reasoning. This release. |
v1.1 taught us the key lesson โ a small model can't memorize every fact, so instead of faking it, v1.11 was retrained to be honest (see the benchmark below).
3 ยท Strengths & limits (honest)
Strengths
- โ Honesty: admits uncertainty instead of hallucinating (HALLU 1โ10, DETAIL 0โ8 vs v1.1)
- โ In-language quality: teaching, long-form, instructions โ all strong across 12 languages
- โ Tool discrimination: answers general knowledge directly, searches only when needed
- โ Multi-perspective on contested history/culture
Limits (be aware)
- โ ๏ธ Multi-step logic/math (LOGIC/ANALYSIS) is weak โ an 8B limit; step-by-step
<think>helps and improves in v1.12 - โ ๏ธ Factual depth is limited (8B) โ for exact facts (law, dates, current data) use RAG/web-search, don't rely on memory
- โ ๏ธ Quantization: use Q5_K_M / Q6_K / Q8_0. Avoid Q4 and below โ heavy quantization made the model hallucinate in our tests.
4 ยท Benchmark (BalkanBench, 13 axes ร 12 languages)
๐ฌ BalkanBench is open โ test any model yourself: https://github.com/olivilo/balkanbench Deterministic scoring (script / language / keywords / numbers). v1.11 (16-bit): 84/156.
Zora leads the field โ beating models 3โ4ร its size, including the current Gemma-4-31B and Qwen3.6-30B:
| Model | Size | Score |
|---|---|---|
| Zora v1.11 | 8B | 84 |
| Gemma-4-31B | 31B | 77 |
| Mistral-24B | 24B | 73 |
| Qwen3.6-30B | 30B | 73 |
| Salamandra | 7B | 66 |
| EuroLLM | 9B | 65 |
| Qwen2.5-32B | 32B | 65 |
| Gemma-2-27B | 27B | 63 |
| Aya | 8B | 61 |
| BgGPT | 7B | 56 |
| YugoGPT | 7B | 35 |
Quantization stays strong โ Q5/Q6/Q8 all remain usable (avoid Q4):
The honesty fix in numbers, and per-axis strengths:
5 ยท Usage
Recommended quant: Q5_K_M (balanced) or Q6_K (near-lossless). Ollama: ollama run olivilo/zora # or: ollama run hf.co/sovasoft/zora-v1.11:Q5_K_M.
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("sovasoft/zora-v1.11")
model = AutoModelForCausalLM.from_pretrained("sovasoft/zora-v1.11", device_map="auto")
๐ Give Zora your own knowledge: see RAG-GUIDE โ connect your own documents (zora_rag.py, Ollama+numpy) or live web search. Zora answers only from the sources and admits when it cannot find the answer.
6 ยท Help Zora grow
Zora's weakest areas are factual depth and some smaller languages. Send us open, licensable sources (texts, corpora, glossaries) in any of the 12 languages โ kontakt@ai.in.rs. See the multilingual manifesto WHY-LANGUAGES on why thinking in your own language matters.
7 ยท Acknowledgements
Zora exists because of open source. We give our formal, heartfelt thanks:
- Above all, to the Qwen team at Alibaba โ for developing and open-sourcing Qwen3 (Apache-2.0), the foundation model Zora is built upon. Without their generosity, Zora would not exist.
- To the platforms and structures that made this possible โ Kaggle, Google Colab, Modal, HuggingFace, Ollama, Unsloth โ for the compute, the tools, and the open infrastructure.
- To the open-source community, for the models, code, and knowledge freely shared with everyone.
- To the people of the Balkans โ whose languages, voices, stories and perspectives are Zora's very heart.
- And to all that is.
ะทะพัะฐ โ the dawn belongs to everyone.
Sovasoft ยท ai.in.rs ยท one to unite them all
- Downloads last month
- 675
5-bit
6-bit
8-bit



Start the llama.cpp server
# Install llama.cpp: brew install llama.cpp# Start a local OpenAI-compatible server: llama serve -hf sovasoft/zora-v1.11: